From patchwork Fri Oct 16 18:11:09 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mark Brown X-Patchwork-Id: 7419851 Return-Path: X-Original-To: patchwork-linux-spi@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork2.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork2.web.kernel.org (Postfix) with ESMTP id 4F590BEEA4 for ; Fri, 16 Oct 2015 18:11:21 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 48454206EB for ; Fri, 16 Oct 2015 18:11:20 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 2751C20742 for ; Fri, 16 Oct 2015 18:11:19 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753614AbbJPSLS (ORCPT ); Fri, 16 Oct 2015 14:11:18 -0400 Received: from mezzanine.sirena.org.uk ([106.187.55.193]:33253 "EHLO mezzanine.sirena.org.uk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752024AbbJPSLQ (ORCPT ); Fri, 16 Oct 2015 14:11:16 -0400 Received: from cpc11-sgyl31-2-0-cust672.sgyl.cable.virginm.net ([94.175.94.161] helo=debutante) by mezzanine.sirena.org.uk with esmtpsa (TLS1.2:RSA_AES_128_CBC_SHA1:128) (Exim 4.80) (envelope-from ) id 1Zn9TE-0000c3-38; Fri, 16 Oct 2015 18:11:12 +0000 Received: from broonie by debutante with local (Exim 4.86) (envelope-from ) id 1Zn9TB-0003EB-9S; Fri, 16 Oct 2015 19:11:09 +0100 From: Mark Brown To: Vignesh R , Mark Brown Cc: linux-spi@vger.kernel.org In-Reply-To: <1444731665-11359-1-git-send-email-vigneshr@ti.com> Message-Id: Date: Fri, 16 Oct 2015 19:11:09 +0100 X-SA-Exim-Connect-IP: 94.175.94.161 X-SA-Exim-Mail-From: broonie@sirena.org.uk X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Spam-Level: X-Spam-Status: No, score=-6.9 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_HI, T_RP_MATCHES_RCVD, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 Subject: Applied "spi: spi-ti-qspi: switch to polling mode for better r/w performance" to the spi tree X-SA-Exim-Version: 4.2.1 (built Mon, 26 Dec 2011 16:24:06 +0000) X-SA-Exim-Scanned: Yes (on mezzanine.sirena.org.uk) Sender: linux-spi-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-spi@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP The patch spi: spi-ti-qspi: switch to polling mode for better r/w performance has been applied to the spi tree at git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted. You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed. If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced. Please add any relevant lists and maintainers to the CCs when replying to this mail. Thanks, Mark From 57c2ecd9bf971946ea0c6ae90a79c90a22159c73 Mon Sep 17 00:00:00 2001 From: Vignesh R Date: Tue, 13 Oct 2015 15:51:05 +0530 Subject: [PATCH] spi: spi-ti-qspi: switch to polling mode for better r/w performance Currently word completion interrupt is fired for transfer of every word(8bit to 128bit in size). This adds a lot of overhead, and decreases r/w throughput. It hardly takes 3us(@48MHz) for 128bit r/w to complete, hence its better to poll on word complete bit to be set in QSPI_SPI_STATUS_REG instead of using interrupts. This increases the throughput by 30% in both read and write case. So, switch to polling mode instead of interrupts to determine completion of word transfer. Signed-off-by: Vignesh R Signed-off-by: Mark Brown --- drivers/spi/spi-ti-qspi.c | 74 +++++++++++++---------------------------------- 1 file changed, 20 insertions(+), 54 deletions(-) diff --git a/drivers/spi/spi-ti-qspi.c b/drivers/spi/spi-ti-qspi.c index aa6d284131e0..89cf0c821524 100644 --- a/drivers/spi/spi-ti-qspi.c +++ b/drivers/spi/spi-ti-qspi.c @@ -39,8 +39,6 @@ struct ti_qspi_regs { }; struct ti_qspi { - struct completion transfer_complete; - /* list synchronization */ struct mutex list_lock; @@ -62,10 +60,6 @@ struct ti_qspi { #define QSPI_PID (0x0) #define QSPI_SYSCONFIG (0x10) -#define QSPI_INTR_STATUS_RAW_SET (0x20) -#define QSPI_INTR_STATUS_ENABLED_CLEAR (0x24) -#define QSPI_INTR_ENABLE_SET_REG (0x28) -#define QSPI_INTR_ENABLE_CLEAR_REG (0x2c) #define QSPI_SPI_CLOCK_CNTRL_REG (0x40) #define QSPI_SPI_DC_REG (0x44) #define QSPI_SPI_CMD_REG (0x48) @@ -97,7 +91,6 @@ struct ti_qspi { #define QSPI_RD_DUAL (3 << 16) #define QSPI_RD_QUAD (7 << 16) #define QSPI_INVAL (4 << 16) -#define QSPI_WC_CMD_INT_EN (1 << 14) #define QSPI_FLEN(n) ((n - 1) << 0) #define QSPI_WLEN_MAX_BITS 128 #define QSPI_WLEN_MAX_BYTES 16 @@ -106,10 +99,6 @@ struct ti_qspi { #define BUSY 0x01 #define WC 0x02 -/* INTERRUPT REGISTER */ -#define QSPI_WC_INT_EN (1 << 1) -#define QSPI_WC_INT_DISABLE (1 << 1) - /* Device Control */ #define QSPI_DD(m, n) (m << (3 + n * 8)) #define QSPI_CKPHA(n) (1 << (2 + n * 8)) @@ -217,6 +206,24 @@ static inline u32 qspi_is_busy(struct ti_qspi *qspi) return stat & BUSY; } +static inline int ti_qspi_poll_wc(struct ti_qspi *qspi) +{ + u32 stat; + unsigned long timeout = jiffies + QSPI_COMPLETION_TIMEOUT; + + do { + stat = ti_qspi_read(qspi, QSPI_SPI_STATUS_REG); + if (stat & WC) + return 0; + cpu_relax(); + } while (time_after(timeout, jiffies)); + + stat = ti_qspi_read(qspi, QSPI_SPI_STATUS_REG); + if (stat & WC) + return 0; + return -ETIMEDOUT; +} + static int qspi_write_msg(struct ti_qspi *qspi, struct spi_transfer *t) { int wlen, count, xfer_len; @@ -275,8 +282,7 @@ static int qspi_write_msg(struct ti_qspi *qspi, struct spi_transfer *t) } ti_qspi_write(qspi, cmd, QSPI_SPI_CMD_REG); - if (!wait_for_completion_timeout(&qspi->transfer_complete, - QSPI_COMPLETION_TIMEOUT)) { + if (ti_qspi_poll_wc(qspi)) { dev_err(qspi->dev, "write timed out\n"); return -ETIMEDOUT; } @@ -315,8 +321,7 @@ static int qspi_read_msg(struct ti_qspi *qspi, struct spi_transfer *t) return -EBUSY; ti_qspi_write(qspi, cmd, QSPI_SPI_CMD_REG); - if (!wait_for_completion_timeout(&qspi->transfer_complete, - QSPI_COMPLETION_TIMEOUT)) { + if (ti_qspi_poll_wc(qspi)) { dev_err(qspi->dev, "read timed out\n"); return -ETIMEDOUT; } @@ -388,9 +393,7 @@ static int ti_qspi_start_transfer_one(struct spi_master *master, qspi->cmd = 0; qspi->cmd |= QSPI_EN_CS(spi->chip_select); qspi->cmd |= QSPI_FLEN(frame_length); - qspi->cmd |= QSPI_WC_CMD_INT_EN; - ti_qspi_write(qspi, QSPI_WC_INT_EN, QSPI_INTR_ENABLE_SET_REG); ti_qspi_write(qspi, qspi->dc, QSPI_SPI_DC_REG); mutex_lock(&qspi->list_lock); @@ -418,31 +421,6 @@ static int ti_qspi_start_transfer_one(struct spi_master *master, return status; } -static irqreturn_t ti_qspi_isr(int irq, void *dev_id) -{ - struct ti_qspi *qspi = dev_id; - u16 int_stat; - u32 stat; - - irqreturn_t ret = IRQ_HANDLED; - - int_stat = ti_qspi_read(qspi, QSPI_INTR_STATUS_ENABLED_CLEAR); - stat = ti_qspi_read(qspi, QSPI_SPI_STATUS_REG); - - if (!int_stat) { - dev_dbg(qspi->dev, "No IRQ triggered\n"); - ret = IRQ_NONE; - goto out; - } - - ti_qspi_write(qspi, QSPI_WC_INT_DISABLE, - QSPI_INTR_STATUS_ENABLED_CLEAR); - if (stat & WC) - complete(&qspi->transfer_complete); -out: - return ret; -} - static int ti_qspi_runtime_resume(struct device *dev) { struct ti_qspi *qspi; @@ -551,22 +529,12 @@ static int ti_qspi_probe(struct platform_device *pdev) } } - ret = devm_request_irq(&pdev->dev, irq, ti_qspi_isr, 0, - dev_name(&pdev->dev), qspi); - if (ret < 0) { - dev_err(&pdev->dev, "Failed to register ISR for IRQ %d\n", - irq); - goto free_master; - } - qspi->fclk = devm_clk_get(&pdev->dev, "fck"); if (IS_ERR(qspi->fclk)) { ret = PTR_ERR(qspi->fclk); dev_err(&pdev->dev, "could not get clk: %d\n", ret); } - init_completion(&qspi->transfer_complete); - pm_runtime_use_autosuspend(&pdev->dev); pm_runtime_set_autosuspend_delay(&pdev->dev, QSPI_AUTOSUSPEND_TIMEOUT); pm_runtime_enable(&pdev->dev); @@ -596,8 +564,6 @@ static int ti_qspi_remove(struct platform_device *pdev) return ret; } - ti_qspi_write(qspi, QSPI_WC_INT_DISABLE, QSPI_INTR_ENABLE_CLEAR_REG); - pm_runtime_put(qspi->dev); pm_runtime_disable(&pdev->dev);